Influence of Near Real-Time Green Vegetation Fraction Data on the Numerical Weather Prediction in WRF over North China

Green vegetation fraction (GVF) has a prominent influence on the partitioning of surface sensible and latent heat fluxes in numerical weather prediction models. However, the multi-year monthly GVF climatology, which is the most commonly-used representation of vegetation states in models, has limited ability to capture the real-time vegetation status. In our study, a near real-time (NRT) GVF dataset generated from 8-day composite of the normalized difference vegetation index (NDVI) is compared with the 10-year averaged monthly GVF provided by the Weather Research and Forecasting (WRF) model. We examine the annual and inter-annual variability of the GVF over North China in details. Many differences of the GVF between the two datasets are found over the dryland cropland and grassland areas. Two experiments using different GVF datasets are performed to assess the impact of the GVF on the forecasts of screen-level temperature and humidity for one year. The results show that using the NRT GVF can lead to a widespread reduction of 2-m temperature in the order of 0.5 ℃, and an increase of 2-m humidity during the warm season. An evaluation against in-situ observations displays an overall positive impact on the near surface parameter forecasts. Over the dryland cropland and grassland areas, a quantitative validation shows that the root mean square errors of 24-h forecasts decline by 9%, 10% and 6% for 2-m temperature, 2-m specific humidity and 10-m wind speed, respectively, in May of 2012. Our study demonstrates that the NRT GVF can provide a more realistic representation of vegetation state which in turn helps to improve the short-range forecasts in the arid and semiarid regions of North China.